Sorry about the delay, I can only post during breaks and when I remember to at home.

Okay, general structure: my (probably overbuilt) serial ISR receives characters and puts them into a buffer (a simple array) accessed only by itself. When the message composed of those characters is complete, it clears its message buffer and puts a unique identifier (a single character) representing that message into the cache data structure discussed earlier. That cache is then accessible by the main program code, which changes its behavior based on what messages are present in the cache.

When the cache is searched, it simply returns the first entry matching the search criteria. That entry is then marked as "old" data, and is eventually overwritten.

Did that clarify matters? I know I should probably put more of the serial ISR's code into the main loop, but it seems pretty happy right now, and is nowhere near long enough to run into its next invocation.

-Bob